Knowledge Graph Construction and Digital Twin Modeling Integrating Multi-modal Data

Abstract

Supply chain analytics is focusing more and more attention on industrial logistics. Optimal allocation of assets is severely restricted by the timing inconsistency and geographical instability of industrial logistics assets brought by unpredictability and variability. Unnecessarily lengthy drive distances and lengthy delays are caused by the incapacity to acquire and utilize industrial logistics asset spatial-and-temporal (ST) qualities rationally, which affects the processes' potential to operate sustainably. Thus, for efficient usage of resources in industrial logistics, a novel machine learning (ML)-assisted knowledge graph (ML-KG) design is provided in this work. For evaluating the multi-modal data produced by widely deployed IoT devices, a novel customizable diversified kernelized support vector machine (CDK-SVM) approach is also suggested. After establishing the suggested ML-KG framework for ST integrity in the representation of the digital twin version, links are carried out and reasoning is performed using job data related to industrial logistics. The graph mechanism effectively distributes the industrial logistics resources. Lastly, the outcome proves that the suggested process is successful in the distribution of industrial logistics assets.